This article is part of a VB special issue. Read the full series: AI and the future of health care
Among the many transformations accelerated by COVID-19, health care ranks at the top of the list. An industry that had been changing at a plodding pace before 2020 has been forced to rapidly embrace advances like telemedicine and health chatbots on a far greater scale to navigate the crisis.
As health care providers adopt these tools, they are receiving a wealth of new patient data that is creating new challenges and opportunities. On the front lines between patients and doctors, the companies driving these products are betting that they are part of a broader revolution that will place data at the heart of everyday treatment.
“We call it digital primary care,” said Nick Desai, CEO of telemedicine platform Heal. “There is still an irreplaceable value to the human-doctor patient interaction. What we want to do is give doctors data-driven decision support.”
Data-driven medicine
Health care was already facing pressure to reinvent itself before the pandemic. A number of trends — such as population growth, longer lifespans, more complicated health issues, and doctor shortages — were among the factors contributing to higher health care costs and strains on the system.
At the same time, a number of digital trends had begun to collide. These include telemedicine platforms, connected health and fitness-monitoring gadgets, and chatbots, which had all steadily increased the amount of digitized health data being produced.
Telemedicine and chatbots got a nice boost when the Centers for Medicare & Medicaid Services expanded reimbursement for remote services such as telemedicine in 2019.
A Stanford University 2020 report published before the pandemic explored the rise of the data-driven physician. Among the factors that seemed to help this trend was the 21st Century Cures Act, passed and signed into law in December 2016. The law set out new data-sharing rules for Electronic Health Record (EHR) systems.
However, it was only last year that the U.S. Office of Management and Budget (OMB) finished defining the rules that would expand patient access to medical records, establish data standards, and enable more interoperability between EHR systems.
“For an industry that has long struggled with low levels of information sharing and poor interoperability across its technology systems, in 2020 we expect to see the final rules create a seismic shift in how health care stakeholders share and interact with digital medical records,” the Stanford report reads. “The rise of the Data-Driven Physician is a sign that the entire health care market is now grappling with the practical application of data and new technologies.”
Then came the pandemic.
Digital health
Even with this forward momentum, many in the medical community were reluctant to embrace these tools. But with the onset of the pandemic, opposition melted away as hospitals became either overwhelmed or simply unsafe to visit.
Hospitals increasingly turned to companies like U.K.-based Babylon Health, which offers services such as video consultations and the ability to report illnesses to providers. The company saw usage soar at the onset of the pandemic, and in May 2020 it launched in its first U.S. market. Sweden’s video consultation platform Kry also launched in the U.S. last spring to address surging demand.
Doctolib, a Paris-based company that offered online booking for medical appointments in France and Germany, had just launched its video feature before the pandemic took hold. Doctolib saw the number of daily video consultations jump from 1,000 pre-pandemic to 100,000 in the first few months of the outbreak. The French government has now authorized it to be one of the main platforms for booking COVID-19 vaccination appointments.
After years of gradual progress, telemedicine and chatbots became overnight successes during the pandemic. According to the recent State of Healthcare report from research firm CB Insights, telehealth (which includes telemedicine and chatbots) became a centerpiece of executive discussions during earnings calls as companies considered how to provide services to employees.
And funding for telemedicine startups soared.
Chatbots
When it comes to chatbots, Ada Health’s combination of artificial intelligence and human doctors had made it a rising star even before the coronavirus began its global spread. The company had spent years developing a platform that allowed patients to input their symptoms so the AI could sort through its databases and either give responses or make a referral to a doctor.
Anyone can download the Ada app for free. To ascertain their symptoms, users are asked a series of questions Ada’s algorithm personalizes based on the responses from each user. The app then suggests possible health issues and proposes next steps, such as making an appointment or going to an emergency room. The app replaces the often tedious work of taking a patient’s history, which can be a big time-saver for doctors and nurses. Ada’s revenue comes via partnerships with health providers who integrate Ada into their early screening systems.
According to Ada cofounder and chief medical officer Dr. Claire Novorol, the company’s consumer app now has 11 million users. While Ada had already seen rapid growth prior to 2020, last year it enabled 5.5 million assessments, or about 25% of all assessments since its app launched in 2016.
Novorol said that during the first phase of the pandemic, when users needed more trustworthy health advice, Ada launched a dedicated COVID-19 assessment and screener to support individuals and health organizations. This screener has since been adopted and integrated into health organizations around the world.
According to Novorol, increased adoption is generating more transparent and consent-driven health data collection and finding ways to share that data will improve digital health services, as well as overall medical quality. This includes capturing data from a wider range of people who might not typically go to a physician.
Novorol said Ada’s access to global aggregated and pseudonymized health data holds the potential to unlock real-time insights that provide additional breadth and depth to treatments.
“In the long-term, not only can health data improve public health and medical quality, but it also has significant potential when it comes to personalization in health care,” Novorol said. “I believe personalized, tailored experiences will be essential to the future of health care — and data will be a key part of that.”
Telemedicine
Heal’s Desai is also bullish about the potential for all this data to drastically improve health care.
The company’s telemedicine platform was initially designed to allow doctors to speak with patients from home. The theory was that seeing patients in their normal setting would be more convenient and give doctors insight into any home conditions that might impact a patient’s health.
Desai outlined four levels of data that can potentially impact health care, with Heal currently delivering the first three. The first level is real-time data that can be provided by simple actions, like a parent taking their kid’s temperature and then sending it directly to the doctor via Heal’s service.
The second level is continuous monitoring of patients via those aforementioned connected devices. Along the way, Heal has developed a suite of tools that allow doctors to continually monitor chronic patients from a distance, including factors like blood pressure, blood sugar, heart rate, and pulse.
That allows physicians to monitor trends in patients’ health status, rather than recording occasional data or relying on patient reporting. Those trends are more powerful for diagnosing a patient because it’s hard to know if a single measure is typical or not. In this case, the doctor can take corrective actions when the trend line seems troubling and more urgent interventions when something seems acute.
“If the doctor knows how your blood pressure’s been doing over the last month, or how your blood sugar has been doing over the last month, that’s very helpful to the doctor to make a more accurate diagnosis,” Desai said. “An average patient is not a good historian of their own health. This way, we keep them out of the hospital, but we’re using that data to more proactively deliver care.”
The third level is looking at the totality of all the data being captured from a patient. This allows for more contextual decisions by looking at a wide range of factors and how they are impacting each other.
However, it’s the fourth level that has Desai particularly excited. The company is currently working with university researchers to develop predictive medicine. This work involves trying to identify what data is useful, how to process it, and what conclusions can actually be made. He estimates such services are at least 12 to 15 months away.
The company is proceeding cautiously because the stakes are enormous. “The key is having it be absolutely accurate enough that the machine’s trend lines are indicative of reality,” Desai said. “Because the moment you make decisions based on machines, they’ve got to be good decisions.”
Even if the company cracks the formula, other hurdles remain. If a doctor can say with a high degree of certainty that a patient will develop a severe illness later in life, it might make sense to consider a preventive procedure. But while that decision might make sense at the time, it could lead to regret later if a treatment or cure for that same illness is developed many years later.
“Those are the kinds of things at an ethical level and at a practical level and at a cost level that become factors,” Desai said. “What is the insurance company willing to pay for the level of knowledge? It’s not just the science that has to advance, it’s also the business of health care, the insurance of health, ethical decisions, therapeutics, and treatment.”
But he added: “This is the holy grail. That machine-driven decision support, that’s the future for us.”
